(1) Field of the Invention
The present invention relates to a structured data management system for assisting in generating structured electronic data and a computer-readable recording medium storing a structured data management program, and more particularly to a structured data management system for assisting in generating electronic data for use in variable non-routine services and a computer-readable recording medium storing a structured data management program.
(2) Description of the Related Art
For handling electronic data such as documents, it has been customary to structure the data according to the logical structure of the contents of the data because it would be difficult to make electronic reuse of unstructured document data. For example, it would require a complex analytic process to extract part of contents from an unstructured document to generate a header or refer to another document.
When electronic data is structured and a system is developed for handling the structured electronic data, if the structured electronic data is fixed in data structure, then the system needs to be changed each time the data structure is changed. If a new system should be able to use old data, it has been the common practice to either convert the structure of the old data into a new structure compatible with the new system or design the new system so as to be able to handle both structures of new and old data. One example of such a system and data conflict is often experienced between an HTML (HyperText Markup Language) and a web browser. For example, tags that have been used by a previous HTML are not used by a new HTML, and there is a web browser which defines unique tags of its own.
The practice of changing a system each time the structure of data is changed requires a large expenditure of time and labor at the time the structure of data is changed. Services practiced in companies change from time to time, and the structure of data handled in those services needs to be changed as the services change. In order to be able to change a data structure freely, there has been proposed a means for defining a logical structure to which each item of data is to adapt itself. One example of such means will be described below.
For processing documents, SGML (Standard Generalized Markup Language)/XML (extensible Markup Language) can be used. SGML is a standard for describing documents as stipulated in IS08879, JIS X4151. XML is a subset of the SGML standard as established by the SGML working group of World Wide Web Consortium. SGML and XML provide a description process for defining a logical structure for documents, known as DTD (Document Type Definition). According to SGML/XML, since the logical structure and layout of a document are separate from each other and the logical structure is clearly indicated by tags in the document, it is easy to process the document with a computer and manage a large amount of data.
SQL (Structured Query Language) can be used to process relational databases. SQL, which is stipulated in JIS X3005, is a database language that is used to define and query relational databases. A schema for a relational database is defined by SQL. The schema is used to define a structure and corresponds to a document type definition for a structured document.
With the process available for defining a logical structure for data, it is possible to electronically process a large amount of data and vary the logical structure of data in response to ever-changing demands.
However, even when SGML which provides the process of defining a logical structure for data is used, if a logical structure is defined by a person and contents of a document are generated by another person, then a problem arises in that semantics of the document may not be unified. Since many more persons are involved in generating a document than persons who define a logical structure therefor, the above problem occurs frequently. As a result, documents may not be generated as intended by persons who have defined the logical structure.
One solution has been to name elements of SGML with definite words. According to another solution, an explanation about how elements of tags of SGML are to be described is described by a clear sentence. These schemes, however, fail to fully eliminate any vagueness about the words and sentences, and also require a large expenditure of time and labor.
SQL provides types, e.g., a number type, a character string type, and a date type, for defining semantics for columns in the definition of a schema for a relational database, and defines contents of types to be written in columns. However, SQL can define contents only in ranges limited by the available types. For example, SQL cannot define an item which designates detailed contents, e.g., writing only past dates except for holidays.
Still another solution to the problem that semantics of a document may not be unified is to use a DynamicHTML which is a descriptive process for making use of an HTML document by joining HTML documents and scripts to each other. More specifically, a script is embedded in a description of an HTML document thereby to generate part of the HTML document dynamically and generate an interactive HTML document. When scripts that define contents of operation of semantics are embedded in parts where elements are written, only descriptions which are not vague can be accepted.
However, if a plurality of expression forms such as HTML documents and scripts are mixed together in one HTML document, as with the DynamicHTML, then both the HTML documents and the scripts cannot easily be read, and cannot be corrected efficiently. For the purpose of describing a service common to HTML documents, scripts may be written in separate files and may be read from scripts of the HTML documents. Since, however, it is not known where and how those scripts are used in the HTML documents, no maintenance can be performed of scripts unless their effect on all the HTML documents is taken into account. According to the DynamicHTML, because scripts are scattered in HTML documents and related to each other, even if some of the scripts are corrected, the effect of the correction on all the scripts must be considered.